This is a repository copy of Cortical thickness and gyrification patterns in patients with psychogenic non-epileptic seizures.
White Rose Research Online URL for this paper:http://eprints.whiterose.ac.uk/131942/
Version: Accepted Version
Article:
Mcsweeney, M., Reuber, M. orcid.org/0000-0002-4104-6705, Hoggard, N. orcid.org/0000-0002-6447-7639 et al. (1 more author) (2018) Cortical thickness and gyrification patterns in patients with psychogenic non-epileptic seizures. Neuroscience Letters, 678. pp. 124-130.
https://doi.org/10.1016/j.neulet.2018.04.056
[email protected]://eprints.whiterose.ac.uk/
Reuse
Items deposited in White Rose Research Online are protected by copyright, with all rights reserved unless indicated otherwise. They may be downloaded and/or printed for private study, or other acts as permitted by national copyright laws. The publisher or other rights holders may allow further reproduction and re-use of the full text version. This is indicated by the licence information on the White Rose Research Online record for the item.
Takedown
If you consider content in White Rose Research Online to be in breach of UK law, please notify us by emailing [email protected] including the URL of the record and the reason for the withdrawal request.
Accepted Manuscript
Title: Cortical thickness and gyrification patterns in patients
with psychogenic non-epileptic seizures
Authors: Marco Mcsweeney, Markus Reuber, Nigel Hoggard,
Liat Levita
PII: S0304-3940(18)30317-3
DOI: https://doi.org/10.1016/j.neulet.2018.04.056
Reference: NSL 33594
To appear in: Neuroscience Letters
Received date: 23-11-2017
Revised date: 24-4-2018
Accepted date: 29-4-2018
Please cite this article as: Marco Mcsweeney, Markus Reuber, Nigel Hoggard, Liat
Levita, Cortical thickness and gyrification patterns in patients with psychogenic non-
epileptic seizures, Neuroscience Letters https://doi.org/10.1016/j.neulet.2018.04.056
This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.
Mcsweeney et al.
1
Title
Cortical thickness and gyrification patterns in patients with psychogenic non-epileptic seizures
Authors Marco Mcsweeney 1*, Markus Reuber 2, Nigel Hoggard 3, Liat Levita 1
Institutional Affiliations 1Department of Psychology, University of Sheffield, Cathedral Court, 1 Vicar Lane, Sheffield, S1 2LT, United Kingdom. E-mail: [email protected] 2Academic Neurology Unit, University of Sheffield, Royal Hallamshire Hospital, Glossop Road, Sheffield, S10 2JF, United Kingdom. 3Academic Unit of Radiology, University of Sheffield, Royal Hallamshire Hospital, Glossop Road, Sheffield, S10 2JF, United Kingdom. * Corresponding author: Marco Mcsweeney. Article counts Word count (text and references): 5108 Abstract: 246 References: 39 Tables: 3 Figures: 1 Supplementary Information: Table 1 & Table 2 Highlights
Our findings partly corroborate but also differ from two previously published brain
morphometry studies in PNES.
These findings point to a highly heterogeneous biopsychosocial disorder in terms of
phenomenology, outcome and presumably aetiology.
The results of this study contradict highly reductionist views and dualistic approaches
to our understanding of PNES.
ACCEPTED MANUSCRIP
T
Mcsweeney et al.
2
Abstract Psychogenic non-epileptic seizures (PNES) are often viewed as manifestations of altered
motor and sensory function resulting from psychological responses to adverse experiences.
Yet many patients and non-expert healthcare professionals find it difficult to understand how
severe disturbances in normal neurological functioning can solely result from underlying
psychological mechanisms to the exclusion of other physical causes. Perhaps importantly,
recent advances using neuroimaging techniques point to possible structural and functional
correlates in PNES. In an attempt to further our understanding of the neurobiological
correlates of this condition, we compared the brain scans of 20 patients with PNES (14
females, mean age 41.05, range 19 – 62) and 20 age- and gender-matched healthy controls
(14 females, mean age 40.65, range 21 - 61) to investigate group differences for cortical
thickness and gyrification patterns using FreeSurfer. Compared to controls, patients
with PNES showed cortical thickness increases in motor, sensory and occipital areas as well
as cortical thickness decreases in temporal and frontal brain regions. In addition, we observed
age-related changes in cortical thickness in the right lateral occipital area in PNES. However,
contrary to our prediction that atypical gyrification may be present, we did not find any
evidence of abnormalities on a measure thought to reflect prenatal and early childhood
cortical development and organization. Nor did we find significant correlations between
cortical thickness results and clinical features. These findings partly corroborate, but also
differ from previous morphometric studies in PNES. These inconsistencies likely reflect the
aetiology and phenomenological heterogeneity of PNES.
Keywords Cortical thickness; FreeSurfer; Local gyrification index; Psychogenic non-epileptic seizures;
Structural magnetic resonance imaging
ACCEPTED MANUSCRIP
T
Mcsweeney et al.
3
1. Introduction
Psychogenic non-epileptic seizures (PNES) are characterized by seizures which
superficially resemble epileptic seizures but in which seizure-like episodes are thought to
result from underlying psychological mechanisms rather than being caused by epileptic
discharges in the brain [1]. In the absence of a clear and easily discernible “organic” cause,
current medical nosologies class PNES as a conversion/somatoform [2] or dissociative
disorder [3]. In light of this, explanations of this diagnosis have largely been rooted in
psychoanalytic or psychological accounts [4], often characterizing these disorders as
medically unexplained [5]. While the latter categorization is a diagnosis of convenience
based on a highly reductionist view of what is considered medically explained, the former
accounts reflect a contested, dualistic approach to the understanding of functional
neurological disorders like PNES [6]. However, there is now a growing body of evidence
from structural and functional studies in PNES which suggests that PNES is best understood
as a biopsychosocial disorder, a disorder in which structural and persistent or recurrent
functional changes in the brain may act as predisposing or precipitating factors for PNES
[7,8].
The present study was intended to add to this evidence by employing whole-brain
cortical surface morphometric analyses of T1-weighted structural magnetic resonance
imaging (sMRI) brain scans of individuals with PNES and age- and gender-matched healthy
controls. First we examined whether age-related changes in cortical thickness (controlling for
gender) would differ between patients with PNES and controls. This is important because
age-related changes in cortical thickness are well documented [9], and disparity between
groups in this regard would have significant implications for how subsequent group
comparisons of cortical thickness are conducted. Secondly, the present study examined
ACCEPTED MANUSCRIP
T
Mcsweeney et al.
4
whether group differences in cortical thickness (controlling for age and gender) would differ
between PNES patients and controls. Based on the two published morphometric studies in
patients with PNES [10,11], it was predicted that we would see group differences in motor,
frontal and occipital regions in addition to brain regions involved in emotion processing.
In addition to cortical thickness measures, we utilized a local gyrification index (lGI)
measure based on that of Schaer et al. [12]. Because the degree of gyrification (gyral and
sulcal formations) is largely determined early in life (primarily during the third trimester with
additional changes during early childhood) and remains relatively stable from adolescence to
adulthood [13], this sensitive measure is thought to be particularly useful for investigating
aberrant early neurodevelopmental changes, traces of which may be identifiable at any age
[12]. While a later age at onset is more common, PNES manifestations have also been
observed during early childhood [14], and given the link between trauma and PNES [15] and
atypical gyrification patterns previously described in children exposed to maltreatment [16]
and in individuals with panic disorder [17], it was predicted that individuals with PNES
compared to controls may show atypical levels of gyrification. Finally, we conducted
correlational analyses to explore the relationship between cortical thickness and PNES
clinical features in cortical regions that showed increases or decreases in cortical thickness in
PNES patients compared to controls.
2. Methods
2.1. Participants
Fifty-three 3T T1-weighted MRI brain scans of patients with PNES acquired between
2009 and 2016 were retrieved retrospectively from the Radiology Department, Royal
Hallamshire Hospital, United Kingdom. Inclusion of MRI brain scans was based upon (a)
confirmed PNES clinical diagnosis by a Consultant Neurologist at the Royal Hallamshire
ACCEPTED MANUSCRIP
T
Mcsweeney et al.
5
Hospital (b) video-EEG recordings of typical attacks with semiological features of non-
epileptic attacks and no associated electro-encephalographic (EEG) or electrocardiographic
(ECG) changes suggestive of epilepsy (c) minimum age of 16 at the time of the scan. MRI
brain scans were excluded if the patient was (a) likely to have had a mixed seizure disorder
(epilepsy and PNES) or (b) had an MRI brain scan showing clinically significant
abnormalities. From this initial PNES sample, twelve scans were excluded due to possible or
definite co-existing epilepsy. Eight scans were excluded due to lack of video-EEG recordings
showing habitual seizure-like episodes. Based upon visual inspection of the MRI scans, three
scans were excluded due to MRI results showing clinically significant brain abnormalities
(two for hippocampal reductions suggestive of mesial temporal sclerosis, and one 76 year old
with T2 hyperintensities which may have reflected a mini stroke), four scans were excluded
due to blurring of the image, and six scans were excluded due to portions of the brain not
being captured in the field of view. No cases were excluded due to age. In addition, fifty-six
3T T1-weighted MRI brain scans of age- and gender-matched healthy controls were retrieved
retrospectively from an existing database of individuals who had previously volunteered for
brain imaging studies (Radiology Department, Royal Hallamshire Hospital). In total, twenty
patients with a “gold standard” PNES diagnosis and twenty age- and gender-matched healthy
controls were included in the analyses. All of the patients with PNES were right-handed.
Age- and gender-matched healthy controls were free from neurological disease or psychiatric
disorders. The retrospective retrieval of archival data was conducted in accordance with the
guidelines set out by the NHS ethical approval granted by the South West – Exeter Research
ethics committee and carried out in accordance with The Code of Ethics of the World
Medical Association (Declaration of Helsinki) for experiments involving humans.
ACCEPTED MANUSCRIP
T
Mcsweeney et al.
6
2.2. PNES clinical features
Semiology features were categorized as 1) generalized motor seizures: seizures
mainly characterized by tonic, clonic, or dystonic-like generalized movements, 2) akinetic
seizures: seizures mainly characterized by unresponsiveness and the absence of movement
with the exception of minor limb tremors, 3) seizures with subjective symptoms: seizures
were mainly characterized by experiential phenomena reported by the patients during video-
EEG recordings, and 4) focal motor seizures: seizures with focal motor movements [18].
Symptom severity was assessed using a symptom severity scale, in which symptom severity
was based upon the summation of previously described clinical features relevant to PNES: 1)
ictal loss of consciousness 2) ictal incontinence 3) ictal tongue-biting 4) ictal injury 5)
accident and emergency attendance for seizures 6) seizure episodes more than thirty minutes
in duration 7) the continuation of symptoms without periods of remission up to the time of
the MRI brain scan [19]. Age at onset, symptom duration, number of anti-epileptic and anti-
depressant drugs taken at time of MRI, clinical psychiatric diagnosis and reports of trauma
exposure were extracted retrospectively from patients’ medical records. Types of trauma
exposure were documented as 1) sexual abuse 2) physical abuse 3) psychological abuse 4)
loss of child 5) other noted but unspecified traumatic experience 6) suicide attempt.
2.3 Image acquisition and FreeSurfer analyses
All brain MRI T1-weighted volumetric scans were acquired using the same Phillips 3
Tesla scanner at The Royal Hallamshire Hospital, Sheffield, U.K. For all participants, brain
MRI was performed using matrix size 256 x 256, field of view 256, slice thickness 1mm,
voxel size 1x1x1, flip angle 8º, coronal plane. Due to the retrospective nature of this study, it
was not possible to limit the MRI pulse sequence parameters used to exactly the same timings
across all of the subjects. Slight differences in echo time (TE) and repetition time (TR) were
ACCEPTED MANUSCRIP
T
Mcsweeney et al.
7
present in our PNES group (TE 3.75 – 4.86 ms, TR 8.16 – 10.42 ms) compared to controls
(TE 4.80 ms and TR 10.50 ms). However, it is important to note that cortical thickness
measurements remain relatively robust even when different MRI protocols or scanners are
used [20].
MRI-based measurements for each participant were obtained using FreeSurfer version
5.30 (http:// www.surfer.nmr.mgh.harvard.edu). In brief, Freesurfer consists of two
processing streams, a surface-based stream and a volume-based stream. The surface-based
stream constructs models of the white matter/gray matter boundary and the boundary
between the gray matter and cerebralspinal fluid (pial surface) from which cortical thickness
measures are taken as the shortest distance between the two. The volume-based stream
preprocesses MRI volumes and labels subcortical tissue classes allowing for the
representation and measurement of subcortical structures (putamen, hippocampus, amygdala,
ventricles etc.). Both cortical and subcortical labelling is based on a subject-independent atlas
and the subject-specific values. These labels are then morphed onto a common space
(average subject) to achieve a common point of reference for each subject. This coordinate
system can be subsequently used to examine group differences by creating group maps. A
full description of this procedure has been published elsewhere [20,21].
Following surface reconstruction and segmentation, the resulting output was visually
inspected for quality and accuracy. If needed, edits were made to adjust for skull strip errors,
intensity normalization failures (requiring addition of white matter control points), incorrect
white matter segmentation, automated topological fixer errors, and pial surface inaccuracies.
In the PNES group one skull strip error was adjusted. Edits to the pial surface were required
in seven scans, topological defects (holes or handles) adjusted in eleven scans, and control
points added in nine scans (263 in total). In the healthy control group, no skull strip errors
occurred, edits to the pial surface were made in twelve scans, topological defects adjusted in
ACCEPTED MANUSCRIP
T
Mcsweeney et al.
8
thirteen scans, and control points added in eleven scans (277 in total). The recon-all
processing stream was re-run from the appropriate stage to recreate the final surfaces in
brains which required corrections to the initial segmentation. To achieve a common point of
reference for each subject the recon-all –qcache flag was used to smooth and resample the
data onto the FreeSurfer fsaverage (average subject made in MNI305 space). Prior to general
linear model analyses (GLM), cortical thickness maps were smoothed with a 10-mm full-
width at half-maximum (FWHM) Gaussian Kernel. No additional smoothing was applied in
relation to lGI. This is because the final FWHM is a composite of the applied FWHM and the
smoothness inherent in the data and lGI has a lot of inherent smoothness. In effect, this
resulted in the degree of smoothness in our lGI data corresponding to a smoothing kernel of
10-mm. Analysis was run on the University of Sheffield high performance computing cluster
(Iceberg; OS 64-bit Scientific Linux (Redhat); 2 X Intel Ivy bridge E5 2650V2 8-core
processors based nodes).
GLM analyses to assess group differences in age-cortical thickness interactions and
lGI-age interactions was run by implementing the mri_glmfit script, with DODS (different
offset, different slope) as design matrix, diagnosis and gender as discrete factors, and age as
covariate. DODS assumes a different offset but predicts a different effect of ageing in both
groups (different slope). To examine group differences in cortical thickness and lGI
(controlling for age), mri_glmfit was used with DOSS (different offset, same slope) as design
matrix, diagnosis and gender as discrete factors, and age as nuisance factor. DOSS also
allows for group differences to start at different points (different offset) but constrains the
group data to change in a similar way, that is, a similar impact of ageing in both groups (same
slope). DODS/DOSS are specific to FreeSurfer (for a detailed description see
https://surfer.nmr.mgh.harvard.edu/fswiki/DodsDoss).
ACCEPTED MANUSCRIP
T
Mcsweeney et al.
9
All of the vertex-wise group analyses were corrected for multiple comparisons using
mri_glmfit-sim in FreeSurfer, with a cluster forming threshold of 3 (p < 0.001) and cluster-
wise probability set to p <0.05. P values were adjusted for both hemispheres using --s
2spaces flag in order to correct for the full search space. This was repeated for 10,000
iterations to derive the location of cluster sizes under the null hypothesis. Clusters surviving
cluster-wise correction were then superimposed on fsaverage inflated surfaces using tksurfer,
a GUI application available in FreeSurfer.
2.4 Correlation analyses with clinical features
We conducted correlational analyses to investigate the relationship between brain
regions that showed increases or decreases in cortical thickness in patients with PNES
compared to controls with clinical features in patients with PNES (age at onset; duration of
symptoms; symptom severity; number of antiepileptic drugs taken). Regions of interest
(ROIs) based on significant cluster-wise corrected cortical thickness results were manually
drawn on the fsaverage inflated surfaces in tksurfer and subsequently mapped back to each
hemisphere for each subject using the mri_label2label command. The average cortical
thickness for each cluster for each subject was then extracted using the mris_anatomical_stats
command. Correlations were conducted using a bivariate nonparametric correlation
procedure (Spearman’s coefficient) with an alpha of 0.05 and subsequently corrected for
multiple comparisons using Bonferroni correction. Kolmogorov-Smirnov and Shapiro-Wilk
were used to test for normality. All statistical analyses were two-tailed and conducted using
the Statistical Package for Social Sciences (IBM SPSS Statistics for Macintosh, Version 24.
Armonk, NY: IBM Corp.).
3. Results
3.1 Demographics
ACCEPTED MANUSCRIP
T
Mcsweeney et al.
10
In total twenty patients with PNES (14 females, mean age at time of scan 41.05,
standard deviation, SD 12.50, age range 19 - 62) and twenty age- and gender-matched
healthy controls (14 females, mean age at time of scan 40.65, SD 12.40, age range 21 - 61)
were included. The mean age at onset of PNES was 27.80 (SD 11.84, range 9 - 51) with a
mean duration of symptoms in years prior to MRI of 10.18 (SD 13.73, range 0.25 – 50). Nine
patients were taking anti-depressants (Supplementary Table 1) and one patient was taking
antipsychotic medication (Quetiapine). Telemetry data capturing typical attacks was available
for all twenty PNES patients. The mean number of PNES habitual attacks recorded was 2.5
(range 1 – 8, Supplementary Table 1). Based on the video-EEG recordings, 45% of patients
(n = 9) were characterized as having predominantly generalized motor seizures/positive
motor phenomena, 35% of patients as having predominantly akinetic seizures characterized
mainly by blank spells with reduced responsiveness (n = 7), and 20% were characterized as
having predominantly seizures with subjective symptoms only but not loss of awareness (n =
4). None were characterized as having focal motor seizures. In all patients, brain MRI was
visually inspected by an experienced neuroradiologist for signs of pathological brain
abnormalities or brain injury. Three patients showed some abnormality. However, these
abnormalities were not deemed clinically significant to the extent that these changes could
explain their symptoms. Given that they affected the white matter and cerebellum they are
unlikely to have affected our cortical thickness analyses. The same morphological analyses
were run with these three missing in addition to their matched healthy controls. Exclusion or
inclusion of these patients resulted in the same clusters and their order for both the left and
right hemisphere. All other patients had unremarkable brain MRI results. PNES group
characteristics are presented in Table 1. The results of PNES symptom severity scale are
presented in Table 2.
ACCEPTED MANUSCRIP
T
Mcsweeney et al.
11
3.2 Morphological analyses
Age-related changes in cortical thickness were first examined after we had controlled
for gender. This analysis identified a single significant cluster surviving cluster-wise
correction in the right lateral occipital area, where patients with non-epileptic seizures
showed greater decreases in cortical thickness with increasing age compared to controls
(Figure 1A & B, Table 3). Group differences in cortical thickness, controlling for age and
gender, were examined next (Figure 1C, Table 3). Cluster-wise corrected results showed
bilateral structural changes in PNES patients compared to controls, with cortical thickness
increases in the cuneus bilaterally, the left paracentral, and left lingual regions. Decreases in
cortical thickness were observed in PNES patients compared to controls in the inferior frontal
gyrus (pars opercularis) bilaterally, right superior temporal region, and the right medial
orbitfrontal cortex. Analysis of gyrification patterns revealed no significant group differences
surviving cluster-wise correction for age-related lGI while controlling for gender or lGI
group comparisons controlling for gender and age. Due to the number of PNES patients who
had reported trauma exposure (Table 1), additional cortical thickness sub-analyses was
conducted in the PNES group only (N = 20). The exact same statistical analyses (described
above) was conducted for age-cortical thickness interactions (controlling for gender) and
cortical thickness controlling for age and gender. The results of these analyses were non-
significant.
---------Insert Figure 1 here-----------
ACCEPTED MANUSCRIP
T
Mcsweeney et al.
12
3.3 Clinical features
Symptom severity positively correlated with cortical thickness in both the left cuneus
(rs = 0.497, p = 0.02) and right cuneus (rs = 0.451, p = 0.04). However, these correlations
were not significant following Bonferroni correction for multiple comparisons. No other
uncorrected significant correlations were found between cortical thickness results and age at
onset, duration of symptoms, or number of antiepileptic drugs taken (Supplementary Table
2). Due to the small number of patients comprising each PNES subtype, it was not feasible to
conduct additional analyses based on semiology.
4. Discussion
The first finding of this study concerns age-related changes in cortical thickness. We
observed cortical thickness differences between groups in the right lateral occipital area
where, compared to controls, patients with PNES showed greater cortical thickness decreases
with increasing age. This is in keeping with a previous cortical thickness study which found
that, compared to healthy controls, PNES patients showed decreased cortical thickness in this
area of the brain [11].
In the second group-level analysis controlling for age and gender, we found that
patients with PNES showed decreased cortical thickness compared to controls in the right
superior temporal gyrus associated with multisensory integration [22] and the right medial
orbitofrontal cortex associated with emotion processing [23], although the direction of the
differences with regard to the right medial orbitofrontal cortex was the opposite of the
findings in a previous study [11]. However, PNES is highly heterogeneous and therefore, it is
possible that this heterogeneity may be the reason for consistent or inconsistent findings
across studies that use similar methodological approaches. Additionally, the lack of well
ACCEPTED MANUSCRIP
T
Mcsweeney et al.
13
defined and established categorical or dimensional characterizations of specific sub-types of
PNES makes it difficult to interpret differing results across studies. We also observed
decreased cortical thickness in regions associated with response inhibition [24,25], namely
the left and right pars opercularis. Interestingly, Labate et al. [10] found that cortical
thickness in the left pars opercularis in PNES patients’ negatively correlated with dissociation
scores, suggesting that higher dissociation scores were associated with decreases in cortical
thickness in this region of the brain. However, it is difficult to make an equivalent inference
between dissociation and our results, as we were unable to directly measure the tendency to
dissociate in our study.
Increased cortical thickness in PNES patients compared to controls was observed in
the left paracentral lobule, with the significant cluster spanning both the primary motor cortex
and primary somatosensory cortex. The paracentral lobule has been associated with, amongst
other things, the planning, control and execution of motor function [26]. However, this
finding differs in terms of both direction and laterality to the findings of Labate et al. [10],
who observed cortical thickness decreases in the right paracentral lobule in patients with
PNES. Again, differences in group characteristic and PNES heterogeneity may be a plausible
explanation for these inconsistencies. Yet, the role of cortical thickness changes in brain
regions involved in motor function is of significant interest in PNES [10,11]. Cortical
thickness increases in PNES were also observed in occipital regions involved in visual
processing [27,28], namely the cuneus bilaterally and the left lingual gyrus. A recent imaging
study [29] found that increased long-range functional connectivity density of occipital
regions (right calcarine fissure and bilateral lingual gyri) correlated with disease duration in
patients with PNES. The authors proposed that changes in functional connectivity in this
region may reflect long-term hypervigilance and increased sensitivity to external stimuli.
While the present study does provide some support for their findings, we did not find a
ACCEPTED MANUSCRIP
T
Mcsweeney et al.
14
significant correlation between cortical thickness results in occipital regions and duration of
PNES.
Furthermore, no significant correlations surviving correction for multiple comparisons
were found between clinical features in PNES and cortical thickness results, nor did we find
cortical thickness differences between patients who had reported traumatic experiences and
those who had not. However, the lack of significant correlations is not altogether surprising.
A number of previous neuroimaging studies have failed to find any significant relationship
between imaging results and clinical features in this disorder, and those that did reported
inconsistent findings [8]. Perhaps more importantly, clinical features derived from medical
records or indeed self-report measures may not be that reliable, especially if they are applied
cross-sectionally in small studies. We must also consider the possibility that changes in
cortical thickness may reflect other factors not accounted for in the present study [30,31],
especially comorbidities often associated with PNES such as anxiety, depression,
posttraumatic stress or personality disorders [32]. Disorders such as these could play an
aetiological role in PNES on the one hand and be related to changes in cortical thickness on
the other. Therefore, it is not clear whether cortical thickness changes associated with PNES
in our study are indeed responsible for PNES or whether changes in cortical thickness reflect
other factors not necessarily associated with this disorder. This is a critical consideration
which has not been sufficiently addressed by our study, nor indeed most other studies which
implicate structural and/or functional brain changes in PNES [8]. This is due to the high
levels of co-existing psychiatric disorders, small sample sizes and lack of psychiatric controls
free of PNES. However, the high level of psychiatric comorbidity observed in our patient
group is in keeping with that observed in most other studies of this disorder and suggests that
we have studied a typical patient sample [32].
ACCEPTED MANUSCRIP
T
Mcsweeney et al.
15
In addition to looking at cortical thickness, we also conducted a group analysis of
gyrification patterns using lGI. However, contrary to our prediction that gyrification may
differentiate PNES patients from controls, the results suggest that atypical gyrification
patterns may not be a contributor to PNES, at least in our sample. Whereas our study
therefore provides some support for the idea that PNES may represent an adaptive (or
maladaptive) process reflected by plastic structural brain changes in frontal, sensorimotor,
temporal and occipital brain regions, we did not find any evidence of abnormalities on a
measure thought to reflect prenatal and early childhood cortical development and
organization [13]. This finding may be surprising. A range of observations provide indirect
evidence for the relevance of neglect and trauma in early life to the development of PNES
[33-36]. Additionally, atypical gyrification patterns have been observed in major depressive
disorder, bipolar disorder, and schizophrenia [37]. Animal studies provide evidence of life-
long structural changes in the brain, neuro-endocrine and behavioural abnormalities after
neglect / trauma in early life, which could underpin these findings in humans [38,39]. It is
possible that the neglect or trauma which may be relevant to PNES affects individuals after
the developmental phase in which gyrification patterns are determined. In addition, neglect or
trauma in early life are not considered an obligatory precondition to the subsequent
development of PNES, but only an important risk factor [6]. Alternatively, our sample may
have been too small or too heterogeneous to pick up relevant structural abnormalities of early
brain development.
In conclusion, our findings of cortical thickness differences between patients with
PNES and healthy controls partly corroborate, but also differ from, morphometry-based MRI
findings in PNES previously described [10,11]. Possible reasons for these variable findings
may include sample size, anatomical variation, and likely differences in group characteristics
in terms of genetic makeup, medical history, life experiences, semiology, duration of the
ACCEPTED MANUSCRIP
T
Mcsweeney et al.
16
disorder, personality characteristics, and co-existing psychopathology. Nonetheless, the key
take home message is that these findings support the growing body of evidence suggesting
that PNES, rather than being a condition that is medically unexplained, may indeed have
physical substrates in the brain. The results of the current study and previous neuroimaging
studies of PNES have important implications for the way we think about and treat individuals
with PNES and how diagnosis may be better communicated to patients. However,
longitudinal morphometric studies prospectively capturing a wide range of demographic,
developmental and clinical data are needed to better address the role of ageing and whether
changes in cortical thickness represent a predisposition to, or consequence of PNES.
Furthermore, PNES are paroxysmal events, which are difficult to investigate through the use
of sMRI data alone. As such, interictal data provides only part of the picture and future
studies should attempt to map electroencephalography (EEG) ictal data acquired during non-
epileptic events to the underlying structure, connectivity and folding patterns of the cerebral
cortex. This may shed more light on the pathophysiological mechanisms of PNES. Future
studies also need to be large enough and involve relevant control groups to allow a better
distinction between the likely associations of PNES itself and of concurrent psychopathology
and or trauma exposure. The interpretation of such datasets would be greatly aided by a
better categorical or dimensional characterization of PNES, a highly heterogeneous disorder
in terms of phenomenology, outcome, and presumably aetiology.
Conflicts of Interest
All authors declare no conflict of interest.
Acknowledgements
This work was supported by a grant from British Academy/Leverhulme [SG142903]
ACCEPTED MANUSCRIP
T
Mcsweeney et al.
17
References
[1] W.C. Lafrance, M. Reuber, L.H. Goldstein, Management of psychogenic nonepileptic seizures, Epilepsia. 54 (2013) 53–67. doi:10.1111/epi.12106.
[2] American Psychiatric Association, Diagnostic and Statistical Manual of Mental Disorders, fifth ed., Arlington, VA, 2013.
[3] World Health Organization, The ICD-10 classification of mental and behavioural
disorders: clinical descriptions and diagnostic guidelines, Geneva, World Health Organization, 1992.
[4] C.M. Monzoni, R. Duncan, R. Grünewald, M. Reuber, How do neurologists discuss functional symptoms with their patients: A conversation analytic study, J. Psychosom. Res. 71 (2011) 377–383. doi:10.1016/j.jpsychores.2011.09.007.
[5] R.J. Brown, Psychological mechanisms of medically unexplained symptoms: an integrative conceptual model, Psychological bulletin. 130 (2004) 793. doi: 10.1037/0033-2909.130.5.793
[6] R.J. Brown, M. Reuber, Towards an integrative theory of psychogenic non-epileptic seizures (PNES), Clin. Psychol. Rev. 47 (2016) 55–70. doi:10.1016/j.cpr.2016.06.003.
[7] M. Reuber, S. Howlett, A. Khan, R.A. Grünewald, Non-Epileptic Seizures and Other Functional Neurological Symptoms: Predisposing, Precipitating, and Perpetuating Factors, Psychosomatics. 48 (2007) 230–238. doi:10.1176/appi.psy.48.3.230.
[8] M. Mcsweeney, M. Reuber, L. Levita, Neuroimaging studies in patients with psychogenic non-epileptic seizures: A systematic meta-review, NeuroImage Clin. 16 (2017). doi:10.1016/j.nicl.2017.07.025.
[9] D.H. Salat, R.L. Buckner, A.Z. Snyder, D.N. Greve, R.S.R. Desikan, E. Busa, J.C. Morris, A.M. Dale, B. Fischl, Thinning of the cerebral cortex in aging, Cereb. Cortex. 14 (2004) 721–730. doi:10.1093/cercor/bhh032.
[10] A. Labate, A. Cerasa, M. Mula, L. Mumoli, M.C. Gioia, U. Aguglia, A. Quattrone, A. Gambardella, Neuroanatomic correlates of psychogenic nonepileptic seizures: A cortical thickness and VBM study, Epilepsia. 53 (2012) 377–385. doi:10.1111/j.1528-1167.2011.03347.x.
[11] A.J. Ristić, M. Daković, M. Kerr, M. Kovačević, A. Parojčić, D. Sokić, Cortical thickness, surface area and folding in patients with psychogenic nonepileptic seizures, Epilepsy Res. 112 (2015) 84–91. doi:10.1016/j.eplepsyres.2015.02.015.
[12] M. Schaer, M.B. Cuadra, N. Schmansky, B. Fischl, J.-P. Thiran, S. Eliez, How to Measure Cortical Folding from MR Images: a Step-by-Step Tutorial to Compute Local Gyrification Index, J. Vis. Exp. (2012) 1–8. doi:10.3791/3417.
ACCEPTED MANUSCRIP
T
Mcsweeney et al.
18
[13] E. Armstrong, A. Schleicher, H. Omran, K. Zilles, O.V. Institut, H.U. Duesseldorf, D. Duesseldorf, The Ontogeny of Human Gyrification, Cereb. Cortex. 5 (1995) 56–63. doi:10.1093/cercor/5.1.56.
[14] M. Reuber, Psychogenic nonepileptic seizures: Answers and questions, Epilepsy Behav. 12 (2008) 622–635. doi:10.1016/j.yebeh.2007.11.006.
[15] R.J. Brown, M. Reuber, Psychological and psychiatric aspects of psychogenic non-epileptic seizures (PNES): A systematic review, Clin. Psychol. Rev. 45 (2016) 157–182. doi:10.1016/j.cpr.2016.01.003.
[16] P.A. Kelly, E. Viding, G.L. Wallace, M. Schaer, S.A. De Brito, B. Robustelli, E.J. Mccrory, Cortical thickness, surface area, and gyrification abnormalities in children exposed to maltreatment: Neural markers of vulnerability?, Biol. Psychiatry. 74 (2013) 845–852. doi:10.1016/j.biopsych.2013.06.020.
[17] S. Yoon, C.S. Jun, H.S. Jeong, S. Lee, S.M. Lim, J. Ma, E. Ko, H.B. Cho, T.S. Yeum, I.K. Lyoo, Altered cortical gyrification patterns in panic disorder: Deficits andpotential compensation, J. Psychiatr. Res. 47 (2013) 1446–1454. doi:10.1016/j.jpsychires.2013.06.015.
[18] A. Magaudda, A. Laganà, A. Calamuneri, T. Brizzi, C. Scalera, M. Beghi, C.M. Cornaggia, G. Di Rosa, Validation of a novel classification model of psychogenic nonepileptic seizures by video-EEG analysis and a machine learning approach, Epilepsy Behav. 60 (2016) 197–201. doi:10.1016/j.yebeh.2016.03.031.
[19] M. Reuber, A.O. House, R. Pukrop, J. Bauer, C.E. Elger, Somatization, dissociation and general psychopathology in patients with psychogenic non-epileptic seizures, Epilepsy Res. 57 (2003) 159–167. doi:10.1016/j.eplepsyres.2003.11.004.
[20] B. Fischl, A.M. Dale, Measuring the thickness of the human cerebral cortex from magnetic resonance images, 97 (20)(2000) 11050–11055.
[21] B. Fischl, FreeSurfer, Neuroimage. 62 (2012) 774–781. doi:10.1016/j.neuroimage.2012.01.021.
[22] H.-O. Karnath, New insights into the functions of the superior temporal cortex, Nat. Rev. Neurosci. 2 (2001) 568–576. doi:10.1038/35086057.
[23] G. Northoff, Functional Dissociation between Medial and Lateral Prefrontal Cortical Spatiotemporal Activation in Negative and Positive Emotions: A Combined fMRI/MEG Study, Cereb. Cortex. 10 (2000) 93–107. doi:10.1093/cercor/10.1.93.
[24] A.R. Aron, T.W. Robbins, R.A. Poldrack, Inhibition and the right inferior frontal cortex: One decade on, Trends Cogn. Sci. 18 (2014) 177–185. doi:10.1016/j.tics.2013.12.003.
[25] D. Swick, V. Ashley, A.U. Turken, Left inferior frontal gyrus is critical for response inhibition, BMC Neurosci. 9 (2008) 102. doi:10.1186/1471-2202-9-102.
ACCEPTED MANUSCRIP
T
Mcsweeney et al.
19
[26] M.R. Borich, S.M. Brodie, W.A. Gray, S. Ionta, L.A. Boyd, Understanding the role of the primary somatosensory cortex: Opportunities for rehabilitation, Neuropsychologia. 79 (2015) 246–255. doi:10.1016/j.neuropsychologia.2015.07.007.
[27] S. Vanni, T. Tanskanen, M. Seppa, K. Uutela, R. Hari, Coinciding early activation of the human primary visual cortex and anteromedial cuneus, Proc. Natl. Acad. Sci. 98 (2001) 2776–2780. doi:10.1073/pnas.041600898.
[28] E. Macaluso, Modulation of Human Visual Cortex by Crossmodal Spatial Attention, Sci. (New York, NY). 289 (2000) 1206–1208. doi:10.1126/science.289.5482.1206.
[29] J. Ding, D. An, W. Liao, G. Wu, Q. Xu, D. Zhou, H. Chen, Abnormal functional connectivity density in psychogenic non-epileptic seizures, Epilepsy Res. 108 (2014) 1184–1194. doi:10.1016/j.eplepsyres.2014.05.006.
[30] B. Draganski, C. Gaser, G. Kempermann, H.G. Kuhn, J. Winkler, C. Buchel, A. May, Temporal and Spatial Dynamics of Brain Structure Changes during Extensive Learning, J. Neurosci. 26 (2006) 6314–6317. doi:10.1523/JNEUROSCI.4628-05.2006.
[31] R.J. Zatorre, R.. Fields, H. Johansen-Berg, Plasticity in Gray and White鳥: Neuroimaging changes in brain structure during learning, Nat. Neurosci. 15 (2013) 528–536. doi:10.1038/nn.3045.Plasticity.
[32] W. Diprose, F. Sundram, D.B. Menkes, Psychiatric comorbidity in psychogenic nonepileptic seizures compared with epilepsy, Epilepsy Behav. 56 (2016) 123–130. doi:10.1016/j.yebeh.2015.12.037.
[33] N. Holman, A. Kirkby, S. Duncan, R.J. Brown, Adult attachment style and childhood interpersonal trauma in non-epileptic attack disorder, Epilepsy Res. 79 (2008) 84–89. doi:10.1016/j.eplepsyres.2007.12.015.
[34] B. Novakova, S. Howlett, R. Baker, M. Reuber, Emotion processing and psychogenic non-epileptic seizures: A cross-sectional comparison of patients and healthy controls, Seizure. 29 (2015) 4–10. doi:10.1016/j.seizure.2015.03.007.
[35] J. Bewley, P.N. Murphy, J. Mallows, G.A. Baker, Does alexithymia differentiate between patients with nonepileptic seizures, patients with epilepsy, and nonpatient controls?, Epilepsy Behav. 7 (2005) 430–437. doi:10.1016/j.yebeh.2005.06.006.
[36] C.G. Kooiman, S. Van Rees Vellinga, P. Spinhoven, N. Draijer, R.W. Trijsburg, H.G.M. Rooijmans, Childhood Adversities as Risk Factors for Alexithymia and Other Aspects of Affect Dysregulation in Adulthood, Psychother. Psychosom. 73 (2004) 107–116. doi:10.1159/000075542.
[37] B. Cao, B. Mwangi, I.C. Passos, M.-J. Wu, Z. Keser, G.B. Zunta-Soares, D. Xu, K.M. Hasan, J.C. Soares, Lifespan Gyrification Trajectories of Human Brain in Healthy Individuals and Patients with Major Psychiatric Disorders, Sci. Rep. 7 (2017) 511. doi:10.1038/s41598-017-00582-1.
ACCEPTED MANUSCRIP
T
Mcsweeney et al.
20
[38] S.J. Lupien, B.S. McEwen, M.R. Gunnar, C. Heim, Effects of stress throughout the lifespan on the brain, behaviour and cognition, Nat. Rev. Neurosci. 10 (2009) 434–445. doi:10.1038/nrn2639.
[39] F. Cirulli, E. Alleva, The NGF saga: From animal models of psychosocial stress to stress-related psychopathology, Front. Neuroendocrinol. 30 (2009) 379–395. doi:10.1016/j.yfrne.2009.05.002.
ACCEPTED MANUSCRIP
T
Mcsweeney et al.
21
Figure caption (black and white version for print) Figure 1. Whole-brain group-level analysis of cortical thickness differences between PNES patients and age- and gender-matched healthy controls. Results depict significant clusters surviving cluster forming threshold of p<0.001 and cluster-wise correction for multiple comparisons at alpha 0.05. Cortical thickness maps were smoothed using a 10mm full-width at half-maximum (FWHM) Gaussian kernel. A = anterior, P = posterior. (A) Group differences in age-cortical thickness interactions controlling for gender. (B) Scatter plot showing age-related changes in average cortical thickness in mm in PNES (triangle) and in age- and gender-matched healthy controls (HC, circle) for the right lateral occipital cluster. (C) Group differences in cortical thickness controlling for age and gender. Figure caption (colour version for online) Figure 1. Whole-brain group-level analysis of cortical thickness differences between PNES patients and age- and gender-matched healthy controls. Results depict significant clusters surviving cluster forming threshold of p<0.001 and cluster-wise correction for multiple comparisons at alpha 0.05. Cortical thickness maps were smoothed using a 10mm full-width at half-maximum (FWHM) Gaussian kernel. Blue and pale blue indicate decreases in cortical thickness. Orange and red indicate increases in cortical thickness. A = anterior, P = posterior. (A) Group differences in age-cortical thickness interactions controlling for gender. (B) Scatter plot showing age-related changes in average cortical thickness in mm in PNES (blue triangle) and in age- and gender-matched healthy controls (HC, red circle) for the right lateral occipital cluster. (C) Group differences in cortical thickness controlling for age and gender.
ACCEPTED MANUSCRIP
T
Mcsweeney et al.
22
Table 1. PNES group characteristics (n = 20)
Comorbid conditions (n = 16) Number Percentage Range* Depression 13 81.3% 1 - 4 Anxiety 9 56.3% Migraine 5 31.3% PTSD 2 12.5% Panic disorder 2 12.5% Agoraphobia 2 12.5% OCD 1 6.3% Fibromyalgia 1 6.3% Number of AEDs taken at time of MRI (n = 12) Number Percentage One 6 50% Two 5 42% Three 1 8% Types of traumatic experiences (n = 10) Number Percentage Range* Sexual abuse 4 40% 1 - 5 Physical abuse 3 30% Psychological abuse 2 20% Loss of child (miscarriage, cot death, other) 5 50% Suicide attempt 3 30% Other trauma unspecified 3 30% Other features (n = 20) Number Percentage History of head injury 1 5% Positive family history of epilepsy 0 0% PNES = psychogenic non-epileptic seizures; PTSD = post traumatic stress disorder; OCD = obsessive compulsive disorder; AEDs = anti-epileptic drugs; MRI = magnetic resonance imaging; *Range refers to the minimum and maximum number of instances i.e. some patients had more than one psychiatric comorbid condition and some patients had been exposed to more than one traumatic event.
Table 2. Results of symptom severity scale in PNES patients (n = 20)
Item Number Percentage
Ictal loss of consciousness 17 85% Ictal incontinence 5 25% Ictal tongue-biting 3 15% Ictal injury 6 30% A&E attendance for seizure episodes 7 35% Seizure duration > 30 minutes 6 30% Recurrent symptoms without periods of remission 17 85% Group scores for symptom severity scale Mean Standard deviation Range 3.05 1.50 1 - 7 PNES = Psychogenic non-epileptic seizures; A&E = accident and emergency
ACCEPTED MANUSCRIP
T
Mcsweeney et al.
23
Table 3. Significant clusters of cortical thickness difference between PNES patients and age- and gender-matched healthy controls for each hemisphere
Group differences in age-cortical thickness interactions controlling for gender (DODS) Cluster
No. Right hemisphere Max Vtx Max Size
(mm^2) MNIX MNIY MNIZ P cluster
Annotation 1 Lateral occipital -5.425 41725 238.03 43.2 -75.4 -7.2 0.01236
Group differences in cortical thickness controlling for age and gender (DOSS)
Cluster No.
Left Hemisphere Max Vtx Max Size (mm^2)
MNIX MNIY MNIZ P cluster
Annotation 1 Pars opercularis -5.217 47493 530.08 -45.6 16.8 21.2 0.00020 2 Paracentral 4.542 116966 174.58 -8.2 -24.6 62.7 0.04996 3 Cuneus 3.938 112494 268.99 -4.4 -84.2 17.8 0.00699 4 Lingual 3.717 114676 193.57 -10.5 -77.7 -4.9 0.03233
Cluster No.
Right Hemisphere Max Vtx Max Size (mm^2)
MNIX MNIY MNIZ P cluster
Annotation 1 Superior temporal -7.052 17325 271.41 44.8 4.4 -23.5 0.00539 2 Superior temporal -6.300 149014 335.88 64.3 -17.9 0.3 0.00160 3 Pars opercularis -5.754 24955 208.01 37.8 19.2 11.3 0.02484 4 Cuneus 5.013 125855 619.63 5.7 -86.9 11.5 0.00020 5 Medial orbitofrontal -3.788 12607 207.47 9.8 43.7 -7.4 0.02544
Results of significant clusters surviving cluster forming threshold (p < 0.001) and cluster-wise correction for multiple comparisons (alpha = 0.05). DODS = different offset different slope; DOSS = different offset same slope; Max = maximum –log10 (p-value) in the cluster with positive and negative values indicating increases or decreases in cortical thickness in patients with PNES compared to controls; Vtx Max = vertex number of the maximum; MNI = Montreal Neurological Institute; MNIX, MNIY, MNIZ = MNI305 coordinates of maximum; P cluster = cluster-wise probability
ACCEPTED MANUSCRIP
T